Near-Real Time Severe Weather Damage Identification Algorithm for Vegetation: Development and Early Results

The NASA Short-term Prediction Research and Transition (SPoRT) Center has partnered with the National Weather Service (NWS) to provide satellite imagery and satellite derived products to the Damage Assessment Toolkit (DAT) to supplement forecasters when performing damage surveys. Satellite remote sensing data has been used in high-impact case studies to examine and document damaged areas that occur as result of severe weather. Unfortunately, this analysis has been only in a manual and time-consuming way. With the advancements in satellite technology near-real time algorithms should be developed to support surveyors in identifying damaged areas that may not be accessible or visible from the ground.

The algorithm that has been developed here combines a short-term Normalized Difference Vegetation Index (NDVI) change product and land surface temperature (LST) information into a feature based technique. These two products are individually analyzed for anomalies within each product and then inputted through a feature detection filter to determine which areas stood out when compared to the entire image.

The development of the algorithm has used MODIS data from Aqua and Terra and past events in order to tune the algorithm. With MODIS exceeding its mission lifetime, this algorithm was designed so that it can easily ingest the next generation of satellites (Suomi –NPP, GOES-R, JPSS-1) which are currently online or coming online in the next couple of years. Most of the results will highlight the algorithm being used on tornado cases. The algorithm's output will be compared against the official NWS ground surveys. An additional case study highlighting the algorithm's success on identification of hail damage will be presented. While the NWS does not currently perform official surveys on wide spread damaging hail events, this algorithm and the DAT could provide the opportunity to expand surveys to cover hail swaths.